Feedback inhibition is the process by which activity in principal neurons stimulates interneurons to release the inhibitory neurotransmitter GABA onto the active principal neurons to prevent runaway excitation. Failure of feedback inhibition is thus a critical element in most theories of the pathogenesis of seizures. However, the functional anatomy of feedback inhibition in the normal brain and epileptic focus is unknown. Recent developments in optogenetics and multiphoton microscopy have made it possible to address this question directly. Here we propose to combine channelrhodopsin-mediated photoactivation of targeted pyramidal cells with high-speed multiphoton imaging of the responses in interneurons expressing the calcium fluorophore yellow chameleon 3.6. These techniques will allow us to define the location of the interneurons that are activated by the target pyramidal cell. After brain injury, the loss of principal neurons and interneurons is compensated by sprouting of new synaptic connections. Our computer modeling suggests that the circuit complexity engendered by this sprouting leaves the interneurons vulnerable to activity-induced synaptic depression that permits runaway excitation and seizures. We hypothesize therefore that epileptic circuits will be defined by characteristic changes in the anatomy of feedback inhibition: more principal neurons will share the same interneuron feedback networks, and individual interneurons will be activated by a wider anatomical range of pyramidal cells, so that the anatomical complexity of local feedback circuits will be increased in epileptic foci. We will test this hypothesis directly with the new optogenetic and microscopy tools in vitro using chronically epileptic organotypic slice cultures, and in vivo using chronically epileptic animals. This data will provide a critical new insight into the pathophysiology of epilepsy that we have not been able to acquire despite wonderfully detailed electrophysiological and classical anatomical studies. Testing for characteristic circuit alterations in epilepsy will make possible new classes of therapeutic interventions including pharmacological manipulation of activity-dependent depression, as well as preemptive activation of critical circuit elements.